Ind. Eng. Chem. Res. 1990,29, 1431-1435
pure compound is described by a Langmuir isotherm gives KiC," qi0 = qsil KiCio
+
(qSiis the saturation coverage of compound i); thus,
In a multicomponent mixture, the adsorption selectivity may be defined as a I 2= zlC2/(zzC1), and from eq 1 a12=
c20r2/(clorl)
(54
The Gibbs isotherm is formulated on the basis that, at equilibrium, the spreading pressures for all the species are the same: ri = rp In addition, if the individual saturation coverages qs2and qsl are quite the same, a12is easily calculated from eqs 4a and 5a: a12
Kirz/(Kzri)
(64
By using Hildebrand's expression for the activity coefficients (cf. Paludetto et al., 1987),
where $i = ( z i / q s i ) / ( E j z j / q sand j ) Aij are interaction parameters such as Aij = Aii = 0. In the case of a binary mixture, this model reduces to In rl = Al2$2/qS1.With the hypothesis qsl = qs2 = qs, $i is equal to zi. Finally,
Registry No. Pyridine, 110-86-1; 4-picoline, 108-89-4; 4ethylpyridine, 536-75-4; 4-tert-butylpyridine, 3978-81-2; 3-picoline, 108-99-6; 3-ethylpyridine, 536-78-7; 3,5-lutidine, 591-22-0; 2picoline, 109-06-8;2-ethylpyridine, 100-71-0; 2,6-lutidine, 108-48-5; 2,44utidine, 108-47-4; 2,4,&collidine, 64-86-8 3,4lutidine, 583-58-4; 3-ethyl-4-picoline, 529-21-5.
1431
Literature Cited Berg, U.; Gallo, R.; Klatte, G.; Metzger, J. Determination of a new scale of ortho-steric parameters Sa from N-methylation of pyriPerkin Trans. 2 1980, 1351. dines. J. Chem. SOC., Bemstein, T.; Kitaev, L.; Michel, D.; Pfeifer, H.; Fink, P. 13Cand 15N NMR and IR spectroscopic investigations of pyridine adsorbed on silica-gel surfaces. J . Chem. SOC., Faraday Trans I 1982, 78,761. Chumakov, Y. I.; Alyabyeva, M. S.; Kaboulov, B. D. Steric effects in liquid-solid chromatographic analysis of pyridine and its 2-alkylsubstituted derivatives. Chromatographia 1975,8, 242. Duprat, F.; Gassend, R.; Gau, G. Inductive adsorption: a new method of isomer separation. Ind. Eng. Chem. Res. 1988,27, 831. Eisenbeiss, F.; Ehlerding, S.; Wehrli, A.; Huber, J. F. K. Optimization of throughput in preparative column liquid chromatography by column overloading and partial fractionation of the effluent. Chromatographia 1985,20, 657. Gassend, R.; Duprat, F.; Gau, G. Adsorption d'amines sur resine sulfonique. Influence de l'encombrement sterique. Entropie 1986, 129, 43. Gassend, R.; Duprat, F.; Gau, G. Inductive adsorption: a new approach to amine enantiomer resolution. J. Chromatogr. 1987,404, 87. Nikitas, P.; Pappa-Louisi, A. Adsorption isotherms for coadsorption studies from solution. Can. J. Chem. 1986, 64, 328. Paludetto, R.; Storti, G.; Gamba, G.; Carri, S.; Morbidelli, M. On multicomponent adsorption equilibria of xylene mixtures on zeolites. Ind. Eng. Chem. Res. 1987,26, 2250. Rahman, M. A.; Ghosh, A. K. Determination of specific surface area of ferric oxide, alumina, and silica gel powders using adsorption of pyridine from n-heptane solutions. J. Colloid Interface Sci. 1980, 77, 50. Rota, R.; Gamba, G.; Paludetto, R.; Carrl, S.; Morbidelli, M. Generalized statistical model for multicomponent adsorption equilibria on zeolites. Ind. Eng. Chem. Res. 1988, 27, 848. Ruthven, D. M. Correlation, analysis and prediction of adsorption equilibria. Principles of Adsorption and Adsorption Processes; Wiley: New York, 1984; Chapter 4. Snyder, L. R. The role of sample structure. Secondary effects. Principles of adsorption chromatography. The separation of non-ionic compounds; Dekker: New York, 1968; Chapter 11.
Received for review October 18, 1989 Revised manuscript received January 30, 1990 Accepted February 15, 1990
Cold Selective Catalytic Reduction of Nitric Oxide for Flue Gas Applications J. P. Chen, R.T.Yang,* and M. A. Buzanowski Department of Chemical Engineering, State University of New York at Buffalo, Buffalo, New York 14260
J. E.Cichanowicz Coal Combustion Systems Division, Electric Power Research Institute, Palo Alto, California 94303
Commercial selective catalytic reduction (SCR) processes for the reduction of NO to N2 by NH3 require temperatures above 300 "C. Significant SCR activities have been measured for transition metal sulfates as the catalysts, a t temperatures as low as 30 "C. Under commercial SCR throughput and gas composition conditions, a 10% NiS04/A1,03 catalyst exhibited NO, conversion of over 50% in the temperature range 50-120 "C and another activity peak a t 200-250 "C with nearly 80% NO, conversion. The low-temperature SCR activity was attributed to the Brernsted acidity of the surface. The Brernsted acidity, and hence the SCR activity, decreased upon the adsorption of water and increased with the presence of SO2 in the gas phase. The selective catalytic reduction (SCR) process for NO, reduction has already gained commercial acceptance in a number of countries, e.g., Germany and Japan, and is presently under consideration for applications in the U S . The stoichiometry of the SCR reaction is 4NH3 4N0 + 0 2 4N2 + 6Hz0
+
-
The commercial catalysts are V20B-based,for example, 0888-5885/90/2629-1431$02.50/0
V2O5/WO3/TiO2.Other catalysts have also been proven to be efficient under power plant flue gas conditions. Examples of these catalysts are H-form zeolites (Kiovsky et al., 1980; Gerdes et al., 1988), transition metal oxides (Powell and Nobe, 1981), and metal sulfates (Ploeg, 1978; DOE Report, 1988, Mori, et al., 1978). A high temperature, however, is required for all of the above-mentioned catalysts. A typical temperature for these catalysts to be ef0 1990 American Chemical Society
1432 Ind. Eng. Chem. Res., Vol. 29, No. 7, 1990 I
t
0 2 I = Figure 1. Schematic diagram of experimental SCR reactor. (A) Chemiluminescent NO/NO, analyzer. (B) NH3 scrubber. (C) Water vapor generator.
b
fective is 400 "C (Kiovsky et al., 1980; Gerdes et al., 1988; Powell and Nobe, 1981; Ploeg, 1978; DOE Report, 1988; Mori et al., 1978). A recent economic analysis of the SCR process for power plant applications (Robie et al., 1989) indicated that significant savings could be effected if the SCR reaction was carried out at a lower temperature, e.g., below 250 "C. Thus, the goal of this study was to search for a low-temperature SCR catalyst. The generally accepted mechanisms for the SCR reaction over V2OSinvolves an Eley-Rideal type mechanism with the following sequential steps: chemisorption of NH3 on a Brernsted acid site (as evidenced by IR of NH4+ species (Takagi et al., 1977; Rajadhyaksha and Knozinger, 1988)),binding of NO to the chemisorbed NH,, desorption of N2 and H20, followed by oxidation of a surface hydroxyl group to a vanadyl group (Inomata et al., 1982). Previous and recent results (Bond et al., 1986; Gasior et al., 1988; Rajadhyaksha and Knozinger, 1988;Yang et al., 1989) have shown that the SCR activity is directly related to the Br~nstedacidity of the V205catalyst. The transition metal sulfates received our primary attention as candidates for the low-temperature SCR catalysts because they are known to possess Brernsted acidity at temperatures as slow as room temperature (Bond et al., 1986). Nickel sulfate, in particular, has been extensively studied and is one of the strongest solid Brernsted acids (Tanabe, 1970, 1988). Experimental Section Catalytic Preparation. Alumina-supported iron, cobalt, and nickel sulfate catalysts were prepared. The alumina support (Kaiser Chemicals, A-305CS) was in the form of pellets (20-32 mesh) with a BET surface area of 341 m2/g. The catalysts were prepared by incipient wetness using aqueous solutions of sulfate or nitrate salt. The impregnated pellets were dried a t 120 "C for 15 h. The dried catalyst samples were immediately capped and were ready for SCR experiments. The sulfate was 10% (by weight) on a water-free basis. Experimental Apparatus and Rate Measurement. The schematic diagram of the SCR apparatus is shown in Figure 1. The reactor was a quartz tube with a fritted support. The reactor was equipped with a thermocouple well and a gas preheating section. The reactant gas was a mixture simulating the flue gas. The flow rates were controlled by an FM 4575 (Linde Division) mass flow control blending system. The same gas mixture was used for all experiments except as noted otherwise, with the following composition: 500 ppm NO, lo00 ppm NH,, 500 2.2% water vapor (when used), and ppm SOz, 2% 02, balance N2. The space velocity was 7500 h-' (based on room temperature and 1 atm) for all experiments.
01
30
r
,
1
3050 100
,
200
3100
Temperature I 'C
do0
5b0
I
Figure 2. Temperature-programmed reaction of NO with NH3over 10% NiSO4/AlZ0,. NO = SOz = 500 ppm, NH3 = lo00 ppm, O2= 2%, N2 = balance, SV = 7500 h-l, heating rate = 2 OC/min.
The N2 was oxygen-free grade (
\-.__
5 40-
0
L
: x
H20 off
20-
-
o---
~
--
io
0
--20
r--e-----
30
40
50
60
70
Time lmln I
Figure 7. Effect of HzO (2.2%) on SCR over 10% NiSO,/Al2O3. NO = SOz = 500 ppm, NH3 = 1000 ppm, O2= 2%, SV = 7500 h-l, T = 50 "C. Table I. Effects of Ammonia Concentration on NO, Conversiono i NO, conversion, %
NH, concn, ppm 250
500 1000
1500 2000
without H,O vapor 14 21
52 76 88
Figure 8. X-ray diffraction patterns of powder collected from reactor effluent (top) and pure ammonium sulfate (bottom).
with H,O vapor 24 24 20 20 20
"Reaction conditions: NO = 500 ppm, SO, = 500 ppm, 0, = 2 % , N2 balance, temperature = 50 "C, H 2 0 = 2.2%, GHSV (space velocity) = 7500 h-l.
The SCR activity could be restored, however, by removing the water vapor from the reactant mixture, as shown in Figure 7 . This result is, again, consistent with the surface structure illustrated in Figure 5. The Bransted acidity could be recovered by desorption of water. Effects of NH3Concentration of SCR Activity. The effects of NH, concentration on the SCR activity of NiS04/A1203at 50 "C are shown in Table I, under both dry and wet Conditions. Under dry conditions, an increase of NH3 concentration beyond the stoichiometric ratio for the SCR reaction (NH,/NO = 1) further increased the NO, conversion. This result indicated that a significant amount of NH, reacted with SOz. The above result, along with the observation that a whitish powder deposit was formed in the cold parts of the SCR reactor in our laboratory as well as in other laboratories and commercial operations (at high temperatures), prompted us to analyze the deposit. The deposit was analyzed by IR absorption spectroscopy and XRD. Both results identified (NHd2S04as the major compound in the powder deposit. The XRD results are shown in Figure 8. The IR spectra of pure (NH4)2S04and of the deposit were identical, both showing strong bands at 3200 and 1410 cm-' (stretching and bending frequencies, respectively, of the ammonium ion) and at 1100 cm-' (stretching frequency of Sod2-). The cause for the formation of (NH,)SO, was likely the following gas-phase reaction: 2NH3 + HzO + SO, (NH4),S04 The data in Table I also show that under wet conditions, the Bransted acidity was low and the SCR activity was not influenced by NH, concentration. Effects of Oxygen Concentration on SCR Activity and Catalyst Life. Inomata et al. (1982) reported for the SCR over V2O5/TiO2that the activity increases with O2 concentration up to 1%Oz but levels off beyond 1%02. The O2effect on the supported sulfate catalyst was studied from 0.5% to 4% 02,and the NO conversions are given in Table 11. It is seen that the effect of O2 on the SCR
-
injection o f SO2 into
the reactant stream
i
o l
0
1
I
10
20
$0
4b
5b
6b
Time [ min.]
Figure 9. Effect of SOz on SCR conversion on 10% NiSO4/AlZO9. NO = SO, = 500 ppm, NH3 = lo00 ppm, O2= 2%, N2 = balance, SV = 7500 h-l, T = 30 O C . Table 11. Effect of Oxygen Concentration on SCR Activity on 10% NiS0,/Alz030 NO conversion, % 0 concn, ppm after 2 h after 3.5 h 0.5 84.0 1 99.4 78.0 2 99.5 85.0 3 99.3 85.0 4 99.1 81.0
"Reaction conditions: NO = 500 ppm, SOz = 500 ppm, O2 = 2 % , N2 = balance, temperature = 50 O C , GHSV = 7500 h-l.
activity on the sulfate catalyst is similar to that on the V205 catalyst. All results reported here are short-time results, i.e., within 2 h. Table I1 also shows that the activity declines beyond 2 h of reaction, at 50 "C. Life tests at both 30 and 50 "C showed steady declines of the activity. However, a steady activity was maintained a t 100 "C. The experiment at 100 "C was done under the following conditions: NH3 = 2000 ppm, NO = SO2 = 1000 ppm, H20 = 2.2%, O2 = 2%, space velocity (SV) = 5000 h-l. A steady NO conversion of 50% was obtained. Subsequently, the following experiments were performed. (a) Without water in the feed, the conversion declined steadily to -20% in 6 h. (b) With 2.2% H 2 0 in the feed, the conversion returned to 50% and showed no decline in activity in a 31-h run. Effects of SO2 on SCR Activity. The SO2 has a promoting effect on the SCR reaction over V20,-based
I n d . Eng. Chem. R e s . 1990, 29, 1435-1438
catalysts (Yang et al., 1989) and zeolite catalysts (Gerdes et al., 1988). The promoting effect of SO2 on the lowtemperature NiS04/A1203 catalysts was considerably stronger. Figure 9 shows this effect at 30 "C. The NO, conversion increased from 42% before SO2 injection to 98% instantly upon SO2 injection and declined slowly to 82%. The apparent effect of SO2 was to increase the Bronsted acidity, by possibly sharing oxygen with the Ni-0-Ni linkage, thus freeing hydrogen from oxygen. We have reported promising results for using transition metal sulfates as the low-temperature SCR catalysts. The reaction is, however, far from being understood. Further studies are in progress in our laboratory. Registry NO.S02,7446-09-5; NH,, 7664-41-7; NO,, 11104-93-1; NO, 10102-43-9 HzO, 7732-18-5; NiS04,7786-81-4; 02,7782-44-7.
Literature Cited Ben-Dor, L.; Margolith, R. An IR and DTA Study of the Hydrates of Metal Sulfates of Cu2+,Co2+,Ni2+,Cd2+,Mn2+,Zn2+,and M8+. Inorg. Chim. Acta 1967, 1 , 49. Bond, G. C.; Zurita, J. P.; Flamerz, S.; Gellings, P. J.; Bosch, H.; Van Ommen, J. G.; Kip, B. J. Structure and Reactivity of Titania Supported Oxides. Part I, Vanadium Oxide on Titania in the Sub- and Super-monolayer Rejions. Appl. Catal. 1986,22, 361. DOE. Quarterly Technical Progress Report, DOE/PETC/QTR88/1; US.Department of Energy: Pittsburgh, PA, 1988. Gasior, M.; Habor, J.; Machej, T.; Czeppe, T. Mechanisn of the Reaction NO + NH3 on V206Catalysts. J . Mol. C a t d 1988,43, 359-369. Gerdes, W. H.; Lim, C. T.; Szymanski, T. Catalyst for the Reduction of Oxides of Nitrogen. U S . Patent 4,735,927, April 1988. Catalyst for the Reduction of Oxides of Nitrogen. U.S. Patent 4,735,930, 1988, both assigned to Norton Company.
1435
Inomata, M.; Miyamoto, A.; Murakami, Y. Activities of V206!Ti02 and V20,/A1203Catalysts for the Reaction of NO and NH3 in the Presence of 02. Ind. Eng. Chem. Prod. Res. Dev 1982, 21, 424-428. Kiovsky, J. R.; Koradia, P. B.; Lim, C. T. Evaluation of a New Zeolitic Catalyst for NO, Reduction with NH,. Ind. Eng. Chem. Prod. Res. Deu. 1980,19, 218-225. Mori, T.; Takeuchi, M.; Hitomi, 0.;Uno, S.; Imahashi, J.; Nakajima, F. Sulfated Metallic Catalysts. US.Patent 4,107,272, 1978, assigned to Hitachi, Ltd. Ploeg, J. E. G. Copper Sulfate and Ammonia for Simultaneous Removal of Sulfur Oxides and Nitrogen Oxides. US. Patent 4,101,634, 1978, assigned to Shell Oil Company. Powell, D. E.; Nobe, K. Reduction of NO with NH3 on Fe-Cr Oxides Catalyst. Chem. Eng. Commun. 1981,10, 103. Rajadhyaksha, R. A.; Knozinger, H. Ammonia Adsorption on Vanadia Supported on Titania-Silica Catalyst an Infrared Spectroscopic Investigation. Appl. Catal. 1988, 51, 81-92. Robie, C. P.; Ireland, P. A.; Cichanowicz, J. E. Technical Feasibility and Economics of SCR NO, Control in Utility Applications. 1989 Joint Symposium on Stationary Combustion NO, Control, San Francisco, 1989. Takagi, M.; Kawai, T.; Soma, M.; Onish, T.; Tamaru, K. The Mechanism of the Reaction between NO, and NH3 on VzOs in the Presence of Oxygen. J. Catal. 1977,50, 441-446. Tanabe, K. Solid Acids and Bases; Academic Press: New York, 1970. Tanabe, K. Acids and Bases in Catalysis. Proceedings of the 9th International Congress on Catalysis; Chemical Institute of Canada: Ottawa, 1988; Vol. 5, pp 85-113. Yang, R. T.; Chen, J. P.; Buzanowski, M. A.; Cichanowicz, J. E. Catalyst Poisoning in the Selective Catalytic Reduction Reaction. 1989 Joint Symposium on Stationary Combustion NO, Control, San Francisco, 1989. Received f o r review October 5, 1989 Revised manuscript received January 30, 1990 Accepted February 15, 1990
Adsorption of Aromatics in NaY and A1P04-5. Correlation with the Sorbent Properties in Separations D. Barthomeuf* and A. de Mallmann Laboratoire de RBactivitB de Surface et Structure, U R A 1106, CNRS, Universitg Paris V I , 4 Place Jussieu, 75252 Paris CBdex 05, France
The adsorption of benzene and o-xylene is studied by means of infrared spectroscopy on A1PO4-5 and Nay. These aromatics do not interact significantly with A1P04-5, but changes in the packing of these molecules alter the shape of the infrared spectra. On NaY a strong cation-?r electron interaction is observed for the two hydrocarbons and for other aromatics. The selectivity in the separation of C8aromatics by weak adsorption on A1P04-5 (preference for o-xylene) is related to the high cohesive energy of this hydrocarbon. The selectivity of these hydrocarbons by strong adsorption on NaY (preference for m-xylene) is related to the favored interaction of this isomer, which is the most basic, with the acidic Na+ ions. The separation of C8 aromatics by selective adsorption on zeolites and zeolite-type adsorbents has been the subject of a large amount of work published largely in the patent literature (1-9). Selectively obtaining each of the four isomers depends mainly for a given zeolite on its cationic form. Experimental practice has shown that the faujasite structure, which offers a large adsorption capacity, can lead to the selective adsorption of p-xylene for KY-based adsorbents ( I , 2, 10, II), of m-xylene for NaY (6,9), and of ethylbenzene for RbX and CsX (4, 5). The rejection of ethylbenzene, as in the Ebex process, may be achieved with CaY or CaX ( 3 ) . Few sorbents are known to adsorb oxylene preferentially, but this is possible on CSZ-1 or 0888-5885/90/2629-1435$02.50/0
A1P04-5materials (8, 9). Changing only the A1 content or the cation in faujasites to achieve very different, even opposite, selectivities strongly suggests that the zeolite chemical properties may influence the sorbed molecules, i.e., that the sorbate-adsorbent interactions are of importance. Correlations were made between the zeolite acidity and the p-xylene selectivity in Y zeolite (IO) or between the zeolite Sanderson intermediate electronegativity and the selectivity for the different isomers (12). The dependence of the interaction between zeolite and aromatics on sorbent chemical properties and loading was also shown in a detailed study of benzene adsorption on a series of faujasites with A1 content from 86 to 10 per unit cell 0 1990 American Chemical Society